David Boyes wrote:
You know, linux can use serial ports as a console device... So why
hasn't IBM come up with a virtual serial port type of console system
to
use instead? Something like having the console on /dev/ttyS0, and
that
via some z/VM magic, is available on an IP as a port number. Telnet
to
the port, and Linux's getty takes it from there. Or better yet,
through
some z/VM magic, the serial ports could be mapped to another Linux
host's serial ports, say one set up as a console appliance... Then
that
appliance could be configured to allow access to them in a variety of
ways, whether it be by port numbers, account names, ssh key, whatever.
Just thinking out loud, but:
You could construct something like this using PVM fairly trivially,
assuming you were willing to tolerate a 3270 console on the Linuxen (and
that the 3270 console driver in Linux tolerates DIALed terminals coming
and going at the appropriate addresses in the virtual machine) and that
PVM was easily available on IFLs. PVM is astonishingly powerful for
doing stuff like this, and Endicott has fixed the licensing problem on
IFLs due to PVM being necessary for CSE. The PVM link signon protocol is
remarkably simple, and could easily be supported on Linux with a little
work.
You could probably also simulate this using YVETTE without spending any
money -- connect to VM, DIAL YVETTE, then DIAL to the guests as
appropriate. I'll have to try this out later.
Wrt to your idea of a console server appliance, that could be done
fairly easily if you run telnetd on a nonstandard well-known port, and
then configure /etc/ttys to allow root logins on that set of ptys. Cisco
has a nice way of mapping serial links to hostnames in IOS that might be
a interesting model. The console appliance would need some type of
relatively secure solicitor as a login shell... hmm. I think this is
possible to do. How badly do you want it? 8-)
I've been imagining this for a long time now, and just wondered why
IBM
never did it.
I think this is how the Integrated ASCII Console thing is supposed to
work.
It doesn't scale all that well, and with the line-mode support in the VM
TELNET server, it's really not all that necessary -- you can already get
a TTY console connection to use to get the network up, and then you can
get directly to the system in question w/o getting the VM system
involved at all.
No, the entire point is to not have to tolerate 3270 at all, to also not
have a line mode in telnet. The entire point is to have a virtual
console device that is *completely* usable. Have it good for vi, screen
(terminal multiplexer), control codes, TUI's, heck, even zmodem uploads
if we wanted to (*shudder*).
As I was saying for the virtual console appliance, there could be many
configurable items. Simple way would be to have folks telnet to a port
and be dropped right into the associated serial console device. Better
way would be to do that with SSH... not sure you could do ports for that
though. Could use ssh keys, and have the appliance side with a shell
script determine which key was used, then send you to a particular
serial console based on that. Or any other number of ways. The main
thing is to get one linux appliance to be attached somehow via virtual
serial connections to all the other Linux guests, and have those guests
configured to use serial console.
If I had to put up with any 3270 limitations or telnet line mode,
there's no point in even pursuing it, as it would be no better than what
we have now... a broken console.
*Brandon
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
